Time Series Prediction Model of Soil Moisture Based on Wavelet De-Noising

Soil moisture is one of the most important factors which affect crop yields directly. In recent years, although upland field area increased not too much, agricultural water resources utilization increased which has led to a waste of water resources in Sanjiang Plain. In order to solve the above problem, wavelet de-noising theory and time series analysis are adopted to analyze measured data of soil moisture after standardization. Then time series forecasting model of soil moisture based on wavelet de-noising is established to simulate and predict soil moisture in Farm Qixing. The results of precision testing and comparative analysis show that the model has high validity and reliability. The model can reveal the time change regulation of area soil moisture and provide scientific basis for reasonably making upland field irrigation system and making full use of soil water resource in Farm Qixing and even the Sanjiang Plain.